Systematic review reveals AI algorithms show comparable sensitivity and specificity in diagnosing HCC to radiologists, suggesting improved workflow efficiencies.
Key Points
AI algorithms achieve sensitivity ranging from 63.0 to 98.6% with specificity of 82.0–98.6%, outperforming junior radiologists.
In contrast, junior radiologists have sensitivity of 41.2–92.0% and specificity of 72.2–100%, indicating varying performance levels.
This systematic review assessed seven studies published from January 2018 to May 2024, all examining AI versus radiologist performance.
While AI can enhance diagnostic accuracy, careful implementation is crucial to prevent bias and overreliance on technology.